Hello, I would like to make a part-time status indicator using IPUMS CPS’s WKSTAT. The codes of WKSTAT changed in 1994, and I am wondering how I can classfy the codes to make a part-time status indicator consistently before and after 1994.
In the attached table, Method 1 classifies the codes based on whether the worker is usually part-time. Method 2 classifies the codes based on whether the worker is currently part-time. Method 3 classifis based on code number (10s are full-time, and 20s and 40s are part-time).
First, I would like to know whether Method 3 is valid. Especially, I would like to know why only “Part-time for non-economic reasons, usually full-time (code 12)” is classified into 10s, although other part-time status are coded into 20s.
Second, I appreciate if you could give me any thoughts of how I can classify the codes consistently before and after 1994.
Method 3 is the method that is implicitly endorsed by the CPS and is therefore the most comparable over time.
Regarding code=12, it really depends on what question your research is aiming to answer. 12=“Part-time for non-economic reasons, usually full-time” is basically identifying individuals who don’t work full time, but not because they can’t work full time - they are choosing not to. Similarly, 13=“Not at work, usually full-time” identifies individuals who have simply taken some time off of work by choice. Your method 3 aligns with the question the CPS originally intended this question to answer which is: how many people are working full-time, how many are working part-time (but would be working full-time if conditions outside of their control were different), and how many are working part-time by their own choice?
The typical and broad advise we give in these sorts of situations is to use all three methods in your analysis. Make note of the specific differences in the construction of each classification, and if your results don’t change between the methods then you can be even more confident in your results. On the other hand, if you are for some reason restricted to use only one method, then choose the method that best answers your research question.